Assignments are piling up. Time is running short. It seems reasonable to resort to more questionable methods to complete your work on time. After feeding an AI assistant each of your assignments, your workload disappears and all is well — at least for now.

Soon, “your” work is noticed and reported. You find yourself before the Office of Student Conduct and Academic Integrity, and depending on how your hearing goes, you could face severe repercussions. Where did it all go wrong?

Recently, it’s been difficult for many to escape the temptation to use AI assistants like ChatGPT or Microsoft Copilot for school assignments. To avoid cases of academic dishonesty, UT administrators have made efforts to cultivate responsible AI use, while some professors have elected to conduct classes using more traditional methods.

In the fall of 2024, UT adopted the AI-Forward AI-Responsible framework meant to foster ways to improve learning using artificial intelligence by spreading awareness about its limitations. These include privacy and ethical concerns; the output of misleading or false information; and the risk of diminished cognitive skills as a result of cognitive offloading, or the use of external tools to reduce the mental load of completing tasks, according to the framework. 

“When AI hit the scene, we started thinking about (the need) to have a guiding principle,” said Julie Schell, assistant vice provost and director of the Office of Academic Technology. “(The principle) means not being so forward that you’re not thinking about (the) important limitations of AI and not being so restricted that you’re not aware of what’s happening with the implementation of AI.” 

However, this framework did not provide enough guidance for the rapidly advancing world of AI, Schell said. In the spring of 2025, Schell joined a team to develop a new framework called Responsible Adoption of AI Tools for Teaching and Learning.

“We realized that we needed more than just that sort of value proposition of AI-Forward, AI-Responsible,” Schell said.  “So we convened a working group on campus of faculty and staff to help us think about, ‘If we’re going to have AI use on campus, what does responsible use, specific to teaching and learning, look like?’” 

This new framework emphasizes using AI thoughtfully, ethically and transparently in ways that promote human agency, uphold academic integrity and balance innovation with responsibility. 

When approaching responsible AI use, Schell said she recommends utilizing four “fusion skills” that help to foster human learning rather than replace it. 

Schell learned the first three skills from a Harvard Business Review article, she said. These include intelligent interrogation, or carefully crafting prompts to interactively guide the AI, reciprocal apprenticeship, or providing richer context in prompts to strengthen the AI model’s understanding of the topic, and critical discernment, or the evaluation of AI output to check for inaccuracies. She developed the last skill herself, which she calls creative transformation, or reshaping an AI model’s response in one’s own voice. 

“I think using AI responsibly means using it in ways that foster your learning and help you grow,” Schell said.

The advent of AI has changed learning permanently for many. Professors now face the new responsibility of identifying AI-generated assignments without any reliable ways to do so, said Rachel Ozanne, associate professor of instruction in the history department 

“I stopped using timed exams during the pandemic, and I was so happy about it,” Ozanne said. “But now, someone can put a prompt into ChatGPT, (which) can write them a paper in five minutes or less. The types of assignments and assessments that I’ve used in the last five years are perhaps not the most effective because AI makes it very easy for students to get something written that they can turn in and get a passing grade.”

To maintain academic integrity and curb AI misuse, professors like Ozanne have had to turn to simpler methods, such as the restoration of timed, in-person exams and oral presentations. Ozanne’s dismissal of AI in the classroom contradicts the framework’s attempts to embrace it as a tool, but Ozanne said she hasn’t seen anything offered from the University in the way of AI that has benefited her.

“There are times when it feels contradictory to be told to make sure we are holding our students to a strong standard of academic integrity … but then also to embrace AI completely and look at all the great things it’s going to do,” she said. “I feel like we just don’t have enough information yet to know what the long-term view of that’s going to be.”

Still, some students have found uses of AI they view as responsible. Economics sophomore Ashley Quilantang uses it to help her study.

“A lot of the time, I have (AI) clarify terms for me,” she said. “For my economics classes, I can get confused a lot on some concepts, so I have it dumb (them) down, which makes it easier for me to understand. When it generates practice questions for me, it helps me fully grasp the material.”

Still, Quilantang said she has growing concerns about her dependence on AI.

“What I’ve been noticing recently is … sometimes when I’m using AI, I’m wondering, ‘Am I not critically thinking?’” she said. “‘Am I just using this out of habit because I’m so used to opening (ChatGPT) when I’m confused, and instead … maybe I could email my TA, or maybe I could ask a friend?’” 

While Quilantang said she maintains academic integrity with her AI use by not passing AI-generated work as her own, she said there are others who do not.

“I know of people (who find) loopholes around using AI during online tests,” Quilantang said. “Whether that be using AI on their phone during the test or during essays, a lot of people write it using (ChatGPT), and then they (ask it to) humanize, humanize, humanize (the output).”

However, according to Chapter 11 of the University’s rules, using AI in this way could land students in academic trouble. According to records obtained by the Texan, reports of academic dishonesty involving AI have been on an upward trend since the spring of 2023, the first semester after ChatGPT was publicly released.

According to Executive Vice President and Provost William Inboden, in a general email provided to the Texan sent to faculty and staff on Aug. 11, if an instructor suspects a student of academic dishonesty, there is no longer an option for the instructor to deal with the situation alone or without input from administration.

This process, labeled “faculty disposition,” has been repealed, requiring instructors to report all suspect cases of academic misconduct to the Office of Student Conduct and Academic Integrity, implementing its student conduct process.

Despite increased reports of irresponsible AI use in the classroom, Schell said most students care deeply about their learning and about the learning of their peers.

“There are a lot of articles out there that are like, ‘Students are cheating their way through college with AI,’” Schell said. “There are some students that are doing that, but most of the students I talked to at UT are really thinking more along the lines of, ‘How do we make sure that we don’t harm our learning with these tools?’ … I think students inherently want to learn, and they’re here to learn and have positive learning experiences.”